Composites: Part A 31 (2000) 353–359

www.elsevier.com/locate/compositesa

Calcium carbonate filled polyethylene: correlation of hardness and yield

stress
J. Suwanprateeb*
National Metal and Materials Technology Center, National Science and Technology Development Agency, Ministry of Science, Technology and Environment,
Rama VI Road, Bangkok 10400, Thailand
Received 30 November 1998; received in revised form 5 August 1999; accepted 1 September 1999

Abstract
The validity of the correlation between hardness and yield stress of particulate composite materials was investigated. Vickers hardness and
monotonic tensile and compression tests have been carried out on calcium carbonate–polyethylene composites at various testing speeds
ranging from 0.2 to 50 mm min 21. It was found that Tabor’s relation was conceivable when the yield stress was determined from compres-
sion test, but not from tensile test. This is associated with the difference in the role and degree of crack formation between two testing modes.
The correlation was also found to be rate dependent and Tabor’s relation was valid when the deformation rate in hardness and yield
determination tests were comparable. q 2000 Elsevier Science Ltd. All rights reserved.
Keywords: Vickers

1. Introduction constrained by the surrounding unstressed matrix. This

relation is found to be useful since yield stress of mate-
Hardness test, being a relatively simple and nondestruc- rials can be estimated from a nondestructive and simpler
tive technique, has been employed widely and accepted as hardness test. However, this simple correlation is not
one of the useful methods for determining the mechanical universal for all cases. In the case of polymers, it has
properties of all kinds of materials. Hardness has been well been reported to be both applicable and inapplicable.
established for characterising metals and ceramics for many Balta-Calleja et al. reported that microhardness of poly-
years, but less for polymers [1,2]. However, due to its ethylene can be related to tensile yield stress for various
usefulness, this technique has been increasingly employed grades of materials [6,7]. However, Henderson and
for polymer characterisation in the recent years. Apart from Wallace reported that the H=s y ratio of high density
quality assurance test, hardness test has been shown to be polyethylene (HDPE) and crosslinked HDPE was much
able to determine changes in morphology and microstruc- less than three [8]. This discrepancy may be due to the
ture of polymers and polymer blends [3,4]. Hardness value greater complex deformation process in polymeric materials
can also be employed to approximate the yield stress of when compared to metals and also the testing conditions
materials. This was originated by Tabor [5] who found which may be of dissimilarity.
that hardness, H, of metals, which are treated as rigid Particulate composites represent growth sector in indus-
perfectly plastic solids, can be related to the yield stress, try due to their competitiveness in terms of cost and perfor-
s y, by the simple equation mance. Large quantities of this type of composite have been
employed in various fields of applications for example in
H < Cs y household, automotive, aviation, etc. Due to the combina-
tion of more than one material, the properties of particulate
where C is a constant and approximately 3. This implies composite are influenced by many factors such as filler
that the pressure employed in hardness test is higher characteristics, filler content, interfacial adhesion, etc. [9–
than the stress needed to initiate yield in quasi-static 11]. These can cause the behaviour of filled polymer to be
test since the stressed material in hardness test is more complex and different from its unfilled counterpart.
The objective of this study is thus to investigate the validity
* Tel.: 1 66-2-644-8150; fax: 1 66-2-644-8077. of yield stress–hardness relationship for particulate filled
E-mail address: [email protected] (J. Suwanprateeb). polymer composites.
1359-835X/00/$ - see front matter q 2000 Elsevier Science Ltd. All rights reserved.
PII: S1359-835 X( 99)00 076-7
354 J. Suwanprateeb / Composites: Part A 31 (2000) 353–359

Table 1
Composition of composites

Sample Volume fraction of CaCO3

A0 0.00
A10 0.10
A20 0.20
A30 0.30
A40 0.40

2. Materials and methods

2.1. Materials

Materials employed in this study was HDPE grade 7000 F

(melt flow index 0.04 g min 210), obtained from Bangkok
Polyethylene Co., Ltd, in the form of pellets and calcium
carbonate powder grade Omega (average particle size
5.53 mm), supplied by Lime Quality Co., Ltd. The densities
of polyethylene and calcium carbonate are 0.956 and
2.70 Mg m 23, respectively.

2.2. Sample preparation

CaCO3 powder and HDPE pellets were first premixed

mechanically at various fractions and the mixtures were
then fed into a co-rotating twin screw extruder (Betol Fig. 1. Shape of tensile and compression specimens.
BTS40L) producing composites with filler volume fractions
ranging from 0.10 to 0.40 (Table 1). Unfilled polyethylene
527. Yielding was determined from the maximum point
was also processed similarly as a control. The composite
where the slope of load–elongation curve reached zero.
pellets were then powdered by a centrifugal mill through a
The tensile yield stress was the load at the maximum
screen mesh 0.5 mm with the help of liquid nitrogen to
point divided by the initial cross-sectional area. Compres-
prevent oxidation. Four mm thick sheets of these compo-
sion tests were performed to ISO 604. However, in this test
sites were prepared by compression moulding. ISO 527
mode, a clearly observable yield point was not able to be
half-sized tensile specimens were cut from the sheets
obtained, no maximum point was observed. Therefore, the
using a contour cutter. The compression samples conformed
yield stress was taken as the stress where the two tangents to
to ISO 604 were right rectangular block of 10 × 10 × 4 mm 3
the initial and final parts of the load–elongation curve inter-
(Fig. 1). The machined surface was subsequently smoothed
sect. All the tests were carried out at 238C and 60% RH. The
by a fine abrasive paper. The measurement of the actual
measurements were performed at constant crosshead speeds
filler content in the composites after processing was
of 0.2, 5, 20 and 50 mm min 21 for both testing modes.
performed by Thermogravimetric Analyser (Perkin–Elmer
TGA System 7). Small pieces of composites were cut from
the compression moulded plates and then scanned in the 2.4. Hardness test
instrument from room temperature to 6008C at the rate of
208C min 21. The change of weight of the samples were The samples were indented at 238C/60% RH by using
monitored and then analysed by the TGA analysis Hardness Tester (Instron-Wolpert Model 930/25) with a
programme. The analysis reveals that the determined filler Vickers diamond pyramidal indenter having a square base
content in all compositions are very close to the percentage and pyramidal angle of 1368. A load of 1 kg (9.8 N) was
of the incorporated fillers, as shown in Table 2. applied for 15 s. The hardness was calculated from measur-
ing the diagonal lengths (d) of the residual indentation on
2.3. Yield determination test the specimen manually using electronic ruler on the
projected screen where an indentation image was displayed
Yield stresses of materials were determined from a through Zeiss objective lenses of magnification 70 × . Vick-
constant elongation rate test in both tensile and compressive ers hardness number was then calculated using the equation
mode. Both tests were performed on a universal testing
machine (Instron 5583). Dumbbell-shaped tensile speci- F
mens (gauge length 25 mm) were tested according to ISO Hv ˆ 1:854
d2
 J. Suwanprateeb / Composites: Part A 31 (2000) 353–359 355

Table 2 properties of materials are largely determined by flaws and

A comparison between experimental and theoretical values of filler content submicroscopic cracks. These flaws act as weak points in
in various composites
the materials. In contrast, the cracks do not play such an
Volume fraction Experimental Theoretical filler important role in compression test because the applied stres-
of CaCO3 filler content (%) content (%) ses tend to close them rather than open them. Thus,
compression properties tend to be characteristic of the
0.00 0 0
0.10 22 24 pure polymer, while tension tests are more characteristic
0.20 36 41 of the flaws in the material. Calcium carbonate–polyethy-
0.30 54 55 lene composites are composed of rigid fillers dispersed in a
0.40 63 65 ductile polymer matrix. These rigid fillers can act as defects
in the composites if the filler is weak or the interface adhe-
where F is load (N) and d is the average diagonal length sion between fillers and matrix is not strong [9] which is the
(mm). case for calcium carbonate–polyethylene composites in the
present study.
2.5. Fractography The evidence of weak interface can be clearly observed
from the SEM micrographs of the fracture surface of the
Fractured samples of composites were viewed microsco- composites after being tested in tension. The micrographs
pically using scanning electron microscope, SEM (JEOL reveal the clean surface of calcium carbonate particles at the
JSM-5410), at the accelerating voltage of 10 kV. The fracture surface without the presence of polyethylene layer
samples were gold sputtered prior to the observation. (Fig. 4). It means that polyethylene debonded completely
from calcium carbonate particles by breaking the interface.
3. Results and discussion This results in the decrease in tensile yield stress with
increasing filler content which corresponds to the increase
When comparing yield stresses which were determined in stress concentration and possible weak points with the
from different modes, tension and compression, opposite increase in calcium carbonate volume fraction. During
trends were observed. It was found that tensile yield stress tensile loading, the interface between polyethylene and
decreases with increasing filler volume fraction whereas the calcium carbonate is pulled apart. It was reported previously
compressive yield stress increased as the amount of filler that tensile yielding of this type of composite was composed
was increased, Figs. 2 and 3, respectively. In general, tensile of interface and matrix yielding [12]. The interface yielding

Fig. 2. Influence of calcium carbonate volume fraction on tensile yield stress of composites.
356 J. Suwanprateeb / Composites: Part A 31 (2000) 353–359

Fig. 3. Influence of calcium carbonate volume fraction on compressive yield stress of composites.

Fig. 4. Fractography of composites showing the detachment of polyethylene matrix from the calcium carbonate surface; (a) A10; (b) A20; (c) A30; and (d) A40.
 J. Suwanprateeb / Composites: Part A 31 (2000) 353–359 357

Fig. 5. Influence of calcium carbonate volume fraction on Vickers hardness of composites.

was initiated from failure of the interface region where the hardness and yield stress of materials will be in the same
fillers debonded from the matrix initiating microcrack with direction since the penetration of indenter in hardness is
statistically distributed sizes and direction. This debonding related to yielding of materials. Therefore, if the hardness
changed the recoverable elastic region to irrecoverable plas- number is increased, the yield stress of materials should
tic region which satisfied the general definition of yielding. increase similarly. However, this relationship is not obvious
Debonding is normally created at low stress level and then for particulate filled polymers. In this study, it was found
grows bigger leading to the interruption of load transfer to that the correlation between hardness and yield stress for
the matrix. Once the separation is sufficient, the sudden drop particulate filled polymer was possible when yield stress
of the load is observed which is a yielding phenomenon in was determined from compressive test only where compres-
load–elongation curve. When the samples were stressed sive yield stress increases with filler volume fraction as
further beyond yield point and the size of the cracks is similar as the hardness number. In the case of tensile
critically large, these cracks will grow rapidly as successive mode, the yield stress decreases with increasing filler
nucleation and microvoid coalescence leading to the failure content in contrast to the trend of hardness value which
of the whole material at last. increases instead. This discrepancy is obviously caused by
However, this phenomenon does not occur in compres- the response of two-phase system upon loading by different
sive test since the interface is in compressive state. There- stress mode. Although there have been reports on the corre-
fore, although the interface failure has occurred, possibly lation between hardness and tensile yield stress, it is clearly
due to the load-induced movement of calcium carbonate emphasized in this study and emphasize the fact that, more
particles through the matrix, the voids enlargement is correctly, the hardness should be related to compressive
limited since the force will tend to close such voids so the yield stress of materials, especially crack sensitive materi-
catastrophic failure is not initiated as in tensile mode. In als, although in some cases, hardness can be related to
addition, the presence of rigid fillers will also act as a retar- tensile yield stress for convenience.
dant of the deformation which comes from the frictional Not only the mode of applied load which is crucial in
force as the particles are forced to move through the matrix. correlating hardness and yield stress of particulate compo-
The compressive yield stress is thus increased with increas- site, but also the rate of the test. It is well known that poly-
ing amount of rigid fillers. mer is a viscoelastic material and its deformation behaviour
Fig. 5 shows the Vickers hardness values (Hv) of compo- is time dependent. Therefore, different rate of deformation
sites at various filler volume fraction. It is observed that the will cause the difference in properties of polymer [13–15].
hardness of composites nonlinearly increases with increas- Lower deformation rate results in the decrease in yield
ing amount of calcium carbonate. Theoretically, the trend of stress and modulus, but elongation at break increases. This
358 J. Suwanprateeb / Composites: Part A 31 (2000) 353–359

Table 3 According to Tabor’s relation, the hardness value is

H=s y ratio of composites in tension and compression test mode at different approximately three times greater than yield stress of mate-
testing speeds
rials. This is due to the fact that indented material in hard-
Samples Hardness/yield stress ness test is constrained by the surrounding matrix whereas
the compressed material does not experience such hydro-
50 mm min 21 20 mm min 21 5 mm min 21 0.2 mm min 21
static component. Thus, the hardness number is greater than
Tension the stress required to initiate yield. In order to study if such
A0 2.06 2.19 2.43 3.44 relation is valid or not, H=s y ratio in compression and
A10 2.63 2.84 2.95 3.74 tension tests were, thus, calculated and tabulated in Table
A20 3.41 3.58 3.85 4.63 3. In compression mode, it is observed that the H=s y ratio is
A30 4.56 4.61 4.95 5.79
A40 4.96 4.93 5.13 5.83
independent of filler content, but changes with rate of test-
Compression ing. This independence implies that the increase in filler
A0 1.80 1.94 2.38 2.68 content causes the proportional increase in both hardness
A10 1.83 1.95 2.33 2.66 and compressive yield stress of composites. Fig. 6 shows
A20 1.90 2.00 2.42 2.83 that the correlation ratio shifts vertically to higher number as
A30 1.96 2.09 2.46 2.78
A40 2.07 2.25 2.45 2.77
the testing speed decreases. The experimental value
approaches the theoretical value described in Tabor’s equa-
tion when the deformation rate decreases to 2 ×
rate-dependence was also found to be the case for polymer 10 21 mm min21 : This deformation rate is of similar order
based composites [12]. In this study, yield stress of all to the rate of the penetration of the indenter into materials
compositions display rate-dependent behaviour in both test- found in general hardness test, which is normally in the
ing modes with increasing yield stress as the deformation order of 10 21 mm min 21. In contrast, the H=s y ratio in
rate increases. Thus, the correlation between hardness and tensile mode is not constant, but nonlinearly increases
yield stress should be influenced by testing speed in yield with increasing calcium carbonate content (Fig. 7). The
determination test and it was shown previously that hard- experimental H=s y ratio is not equivalent to the Tabor’s
ness of various grade polyethylene was affected by the constant, either lower or greater even at the comparable defor-
deformation rate. The correlation between hardness and mation rate. The increase in the H=s y ratio with filler content in
yield stress was observed if the deformation rate for both tensile mode clearly demonstrates the dependence of tensile
tests were in the similar ranges [7]. This is found to be the yielding on the interfacial debonding. Thus, it is seen here that
case for calcium carbonate–polyethylene composite in the Tabor’s relation is a function of deformation rate and can be
present investigation as well. applied for calcium carbonate–polyethylene composite with

Fig. 6. Influence of testing speed and calcium carbonate volume fraction on the H=s y ratio in compressive mode.
 J. Suwanprateeb / Composites: Part A 31 (2000) 353–359 359

Fig. 7. Influence of testing speed and calcium carbonate volume fraction on the H=s y ratio in tensile mode.

varying filler contents when the deformation is in compres- References

sion mode and deformation rate is comparable to that under
indenter. [1] Askeland DR. The science and engineering of materials. 2. London:
Chapman and Hall, 1984.
[2] Fee AR, Segabache R, Tobolski EL. ASM handbook, Mechanical
4. Conclusions testing, 8. USA: ASM International, 1995. p. 90.
[3] Rueda DR, Martinez-Salazar J, Balta-Calleja FJ. J Mater Sci
The correlation between hardness and yield stress which 1985;20:834.
[4] Balta-Calleja FJ, Fakirov S. Trend Polym Sci 1997;5(8):246.
can be generally observed in many homogeneous materials [5] Tabor D. The hardness of metals. New York: Oxford University Press,
is not straightforward for particulate filled composites, espe- 1951.
cially for weak interfacial adhesion type, since the suscept- [6] Balta-Calleja FJ. Adv Polym Sci 1985;60:119.
ibility of composites under two modes of tests is dissimilar. [7] Balta-Calleja FJ, Giri L, Ward IM, Cansfield DLM. J Mater Sci
The use of hardness number to determine or estimate yield 1995;30:1139.
[8] Anderson PJ, Wallace AL. Polymer 1989;30:2209.
stress of composites is possible but has to be carried out with [9] Demjen Z, Pukanszky B, Nagy J. Composites Part A 1998;29:323.
care. Selection of testing mode and speeds should be consid- [10] Liu CT, Tang B. Composites Part B 1998;29:15.
ered in achieving correct results. [11] Suwanprateeb J, Tanner KE, Turner S, Bonfield W. J Mater Sci:
Mater Med 1995;6:804.
[12] Suwanprateeb J, Tiemprateeb S, Kangwantrakool S, Hemachandra K.
Acknowledgements J Appl Polym Sci 1998;70:1717.
[13] Matz DJ, Guldemond WG, Copper SL. Polym Engng Sci
Bangkok Polyethylene Co., Ltd and Lime Quality Co., 1973;13:300.
[14] Povolo F, Schwartz G, Hermida EB. J Appl Polym Sci 1996;61:109.
Ltd are thanked for the supply of raw materials. S. Kang-
[15] Hamdan S, Swallowe GM. J Mater Sci 1996;31:1415.
wantrakool was acknowledged for help in preparing the
composites and S. Tiemprateeb for carrying out parts of
tensile and compression tests.